Introduction

Septic patients with liver cirrhosis exhibits considerable heterogeneity.

Aims & Methods

The objective of current study was to identify subphenotypes of liver cirrhosis with sepsis among intensive care unit (ICU) patients. A retrospective study was performed based on the Medical Information Mart for Intensive Care (MIMIC-IV 2.0) database. Using unsupervised clustering, we identified distinctive subphenotypes based on 37 clinical variables obtained on admission. Primary outcomes were 28-day mortality. Cox regression was applied to quantify the risk associated with each subphenotype concerning clinical outcomes. A machine learning (ML) prediction model was developed and made available online, capable of identifying these subphenotypes from fundamental clinical variables.

Results

Our study enrolled a total of 1087 septic patients with liver cirrhosis, revealing two unique subphenotypes via k-means clustering algorithm. Subphenotype 1 was assigned to 719 patients (66%) and subphenotype 2 to 368 patients (34%). The subphenotype 2 was characterized by higher levels of hepatic injury, higher SOFA scores (average SOFA scores were 12.5), and worse clinical outcomes (28-day mortality, 49%). Subphenotype 2 showed significantly increased risk of 28-day mortality compared to Subphenotype 1 (hazard ratio [HR] = 2.14 [95%CI 1.69 – 2.71], p < 0.001). The ML model integrated seven variables could predict subphenotypes accurately.

Conclusion

This study differentiated and stratified subphenotypes among septic patients with liver cirrhosis, laying a foundation for more personalized therapeutic strategies. The predictive ML model developed has the potential to bolster clinical decision-making pertaining to these patients within the ICU setting.

Disclosure

The authors declare that they have no conflict of interest.

Introduction

Alcoholic liver cirrhosis (ALC) is a complex condition marked by numerous complications, often leading to a poor prognosis and high mortality rates^​. This study evaluates the clinical characteristics, causes of hospitalization, and factors associated with prolonged hospital stay and in-hospital mortality among patients with alcoholic liver cirrhosis.

Aims & Methods

A retrospective study was conducted on 357 patients diagnosed with alcoholic liver cirrhosis in a tertiary gastroenterology department. The mean age was 57.8 ± 9.8 years, and 85% were male. Data on causes of hospital admission, length of stay, and clinical complications were analyzed. Associations with prolonged hospitalization (defined as >3 days) and in-hospital mortality were assessed using univariate and multivariate statistical analyses.

Results

The overall in-hospital mortality rate was 21.5%. The most common causes of admission included urinary tract infections (57.7%), hepato-renal syndrome (28.2%), digestive hemorrhage (22.6%), acute alcoholic hepatitis (19.3%), spontaneous bacterial peritonitis (8.4%), and pulmonary infections (5.6%). The median hospital stay was 6.6 days.
Variables significantly associated with prolonged hospitalization in univariate analysis included ascites (p<0.0001), spontaneous bacterial peritonitis (p=0.01), hepatic encephalopathy (p<0.0001), urinary tract infections (p<0.0001), and hepato-renal syndrome (p=0.09). In multivariate analysis, only urinary tract infection was independently associated with prolonged hospital stay.
Regarding in-hospital mortality, significant associations in univariate analysis were found with spontaneous bacterial peritonitis (p<0.0001), acute alcoholic hepatitis (p<0.0001), digestive hemorrhage (p<0.0001), and hepato-renal syndrome (p<0.0001). In multivariate analysis, only hepato-renal syndrome remained independently associated with mortality (p<0.0001).

Conclusion

Urinary tract infection is an independent predictor of prolonged hospitalization in patients with alcoholic liver cirrhosis, while hepato-renal syndrome is the only independent predictor of in-hospital mortality. These findings highlight the importance of early recognition and management of infections and renal complications in this high-risk population.

Introduction

Non-invasive elastography techniques have become essential for evaluating hepatic fibrosis and portal hypertension. Although liver stiffness is routinely assessed, spleen stiffness has gained recognition as an emerging marker of portal hypertension. Its application in the intensive care unit (ICU), particularly among patients with chronic liver disease and acute-on-chronic liver failure (ACLF), remains underexplored. This study investigates the utility of Acoustic Radiation Force Impulse (ARFI) imaging for spleen stiffness assessment in ICU patients, focusing on its prognostic relevance and feasibility in a high-acuity setting.

Aims & Methods

We conducted a prospective observational study enrolling 44 ICU patients with established chronic liver disease, with or without ACLF. Spleen stiffness was measured serially using ARFI elastography every 2–4 days during the ICU stay. Clinical data—including liver and renal function parameters, respiratory support settings, and outcome variables—were collected concurrently. The primary aim was to assess the feasibility of serial spleen stiffness measurement in this cohort, identify influencing clinical variables, and explore associations with clinical outcomes.

Results

Spleen stiffness assessment via ARFI was successfully performed in all included patients. Higher spleen stiffness values were significantly associated with renal replacement therapy (p < 0.001) and the need for invasive mechanical ventilation (p < 0.001). An inverse relationship was observed between airway pressures during ventilation and spleen stiffness values (p = 0.017), suggesting potential interplay between mechanical factors and elastographic findings. Spleen stiffness did not correlate with patient age, BMI, or vasopressor requirement. Over time, spleen stiffness showed dynamic changes corresponding to clinical progression or improvement, supporting its potential as a non-invasive surrogate of disease trajectory.

Conclusion

Our findings demonstrate that spleen stiffness measurement via ARFI elastography is both feasible and clinically informative in the intensive care setting for patients with liver cirrhosis. Beyond reflecting portal hypertension, spleen stiffness appears to integrate aspects of hepatic and extrahepatic dysfunction, correlating with organ failure and clinical deterioration. These results underscore its potential as a non-invasive, bedside-accessible biomarker to support dynamic risk stratification and individualized decision-making in critical care hepatology.